There is known a communication system in which two time servers are connected respectively to two networks and a communication device is connected to both the networks (see, e.g., NPL 1).
In NPL 1, Grandmaster is a time server, and Ordinary clock is the communication device. The communication device synchronizes in time with a main time server and switches over to a sub time server when being in an incommunicable state with the main time server. The redundancy of both the communication path and time server increases the reliability of a time synchronization system.
However, in general, a device, a method, and a program for specifying a failure part have not been established. An administrator cannot grasp the abnormal part before restoration. Therefore, a replacement cannot be prepared until he or she specifies the abnormal part at the installation site of a communication system, making restoration work inefficient.
Conventional techniques have a problem that it is not possible to specify the abnormal part.